Discover Nitinguptadu/Object-detection-code- Open Source project

Introduction The original code of Keras version of Faster R-CNN I used was written by yhenon (resource link: GitHub .) He used the PASCAL VOC 2007, 2012, and MS COCO datasets. For me, I just extracted three classes, “Person”, “Car” and “Mobile phone”, from Google’s Open Images Dataset V4. I applied configs different from his work to fit my dataset and I removed unuseful code. Btw, to run this on Google Colab (for free GPU computing up to 12hrs), I compressed all the code into three .ipynb notebooks. Sorry for the messy structure. I wrote my exploring and experiment results for Faster R-CNN in this article in Medium. If you are in China, you cannot directly access Medium. So I make a copy in here. Project Structure Object_Detection_DataPreprocessing.ipynb is the file to extract subdata from Open Images Dataset V4 which includes downloading the images and creating the annotation files for our training. I run this part by my own computer because of no need for GPU computation. frcnn_train_vgg.ipynb is the file to train the model. The configuration and model saved path are inside this file. frcnn_test_vgg.ipynb is the file to test the model with test images and calculate the mAP (mean average precision) for the model.

Faster-R-CNN-for-Open-Images-Dataset-by-Keras

In this Repository I have deployed the faster Rcnn model on flask to deploy it on local server and aslo provided the code for how to use AWS sagamaker to deploy your pre train model on aws

Faster-R-CNN-model-deployment-using-flask-on-local-host

In this challenge, a full stack of EM slices will be used to train machine learning algorithms for the purpose of automatic segmentation of neural structures. The images are representative of actual images in the real-world, containing some noise and small image alignment errors. None of these problems led to any difficulties in the manual labeling of each element in the image stack by an expert human neuroanatomist. The aim of the challenge is to compare and rank the different competing methods based on their pixel and object classification accuracy.

ISBI-Challenge-Segmentation-of-neuronal-structures-in-EM-stacks

kaggle competition iMet Collection 2019 Classify the multi-‐label images classification according to their given label using a ResNet50 . build the model from the scratch

Resent-50-from-scratch-for-multi--label-image-classification

Classify the multi-‐label images classification according to their given label . build the model from the scratch in keras

Multi-label-image-classification-in-keras

yolo v3 in pytorch ( python version 3 ) for image real time image detection and video detection (video formate .avi supported by opencv )

Yolo-V-3-network-from-scratch-in-pytorch

The objective was to predict the chances of passenger’s survival Challenges – Dealing with missing value and Data Cleaning Techniques used One Hot encoding for categorical data and Min Max Scaler for numerical data Model used – Logistic Regression, Decision tree (ID3 algorithm), Random Forest Classifier

Titanic-Data-set-

The labelled dataset below will have images of damaged cars along with labels for various types of damages. Among various categories of damages, The model only 2 damage categories- scratches or spots, and dents . Use various image enhancement and augmentation techniques to improve accuracy of the model.

Object-Detection-using-Fast.ai

All code was taken from Max deGroot's & Ellis Brown's ssd.pytorch repository except the object_detection.py file. However, some modifications were done in order to make this project run on Windows 10 and Python 3.6 with PyTorch 0.4.1

PyTorch-SSD

Faster-R-CNN-model-dockerization

Nitinguptadu

Breast cancer detection

CNN code in pytorch for raw image in Jupiter notebook with accuracy of 96.2%

CNN-code-in-pytorch

key word recognition ( one to nine ) using CNN

key-word-recognition-using-CNN

I also noticed there are lots of images in the data which are specific B&W or only of R/B/G channel. Based on these observations I decided to write the below code to do small changes in images which are from unbalanced classes in training sample ans save them:

Pillow-image-data-augumenatation-using-manullay-preprocessing

Coding-

python basic A to Z

This Model Runs on Free Heroku server .Static page is inside app.py . Model is loaded by keras application

Image-Classifiaction-Using-Resnet-

https-github.com-Nitinguptadu-To-acess-gpu-from-local-computer-in-pytorch-and-keras-

Manullay created Hot one encoding and find unique values form csv

Manually-created-one-hot-encoding-

Gpu GTX 1650 with 4 gb Graphic Ram Cuda version 10 tensorflow-gpu==1.13.1

Keras-Gpu-installization-for-yolo-v3-for-text-detection-

Yolo V 3 network from scratch in pytorch

Yolo-v3-

Gas-sensor-array-under-dynamic-gas-mixtures

Time series data for Ethylene and Methane in air, and Ethylene and CO in air

This repo is for self learning Purpose

CONV2D-with-Tabular-Data-

Multi-Label Image Classification Model in Python (Kearas)

To-acess-gpu-from-local-computer-in-pytorch-and-keras-

Multi-Label-Image-Classification-Model-in-Python

cnn code in pytorch for raw colour image with accuracy of 96.2 %

CNN-code-for-raw-image-

Image-classification-Heroku-

This code is for self learning purpose

github-

Flye

Fine-tune VGG16 Image Classifier with Keras

Fine-tune-VGG16-Image-Classifier-with-Keras-

Autoencoder-cifar10.py-using-keras.ipynb

Fine-tune VGG16 Image Classifier with Keras

Fine-tune-VGG16-Image-Classifier-with-Keras

Tsfresh

Linear Regression , logistic Regression , KNN, K Mean ,DecisionTrees_RandomForest_Classification,Feature Engineering,K fold,PCA,Random_Forest_Regression,RMSE

Machine-learning-A-to-Z

Busigence-Assigment-multilable-image-classifictation-in-keras

multi-‐label classification problem Classify the images according to their given label .build the model from the scratch

Raw-Fashion-image-data-classification-using-cnn

this repo is for self learning purpose

cat-and-Dog-Nitin

converting and resize colour image into gray scale using opencv

Converting-RBG-image-to-gray-scale

Calculate the Screen Time of Actors in any Video (with Python codes) using Convolutional Neural Networks

Calculate-the-Screen-Time-of-Video-Using-CNN

Implementing-Autoencoders-in-Keras-Tutorial

In this tutorial, you’ll learn about autoencoders in deep learning and you will implement a convolutional and denoising autoencoder in Python with Keras. You will work with the NotMNIST alphabet dataset as an example

multi-‐label classification problem

Raw-Fashion-image-data-for-image-classification-using-CNN

Image-classification

Classification of MNIST data set

image classification using pytorch

Pytorch-Basic

Image classification using MNIST Fashion data in pytorch

Pytorch-CNN

Cifar-10-data-set-with-train-accuracy-99-and-test-accuracy-93-

Linear-regresssion

speech_to_text

Multi label image classification problem

Fast.ai

multi-‐label classification problem

Nitinguptadu-Raw-Fashion-image-data-for-image-classification-using-CNN

Dentisty.AI

Please use the following assignment for Data Scientist position: The deadline of Assignment is 3 days. 1. Download the data-set from : https://warwick.ac.uk/fac/sci/dcs/research/tia/glascontest/download/ 2. Augment the data-set to make it efficient against shift, light variation and other noise. 3. Train a transfer learning based classifier to classify images according to benign/malignant given in the Excel file. 4. Train a transfer learning based Semantic segmentation according to the annotation in _anno.BMP file. 5. Upload the assignment on your Github in form of a IPYNB showing results on a few test set images.

Build a RESTful service that extracts expense date from a receipt.Deploy the service on any cloud platform like Heroku/AWS/GCP. The service should contain one API which has the following contract: Request: POST /extract_date Payload: {“base_64_image_content”: <base_64_image_content>} Response: If date is present: {“date”: “YYYY-MM-DD”} If date is not present: {“date”: null}

Build-a-RESTful-service-that-extracts-expense-date-from-a-receipt.

Chest-X-Ray-Images-Pneumonia-

using Fast.ai

Develop a generalized algorithm to detect the brightness of any image. Your algorithm should take an image as input and give a score between (0-10) as output (zero being low bright and 10 being high bright). You will not be provided with data for training.

Develop-a-generalized-algorithm-to-detect-the-brightness-of-any-image

This series will teach you how to use Keras, a neural network API written in Python. Each video focuses on a specific concept and shows how the full implementation is done in code using Keras and Python. We will learn how to preprocess data, organize data for training, validation and testing, build an artificial neural network from scratch, train an artificial neural network, build a convolutional neural network (CNN) and much more!

Keras---Python-Deep-Learning-Neural-Network-API

Faster-Rcnn-docker-

Let's implement resnet from scratch in pytorch

Resnet-50

a new framework for estimating generative models via an adversarial process, in which we simultaneously train two models: a generative model G that captures the data distribution, and a discriminative model D that estimates the probability that a sample came from the training data rather than G. The training procedure for G is to maximize the probability of D making a mistake. This framework corresponds to a minimax two-player game. In the space of arbitrary functions G and D, a unique solution exists, with G recovering the training data distribution and D equal to 1/2 everywhere. In the case where G and D are defined by multilayer perceptrons, the entire system can be trained with backpropagation. There is no need for any Markov chains or unrolled approximate inference networks during either training or generation of samples. Experiments demonstrate the potential of the framework through qualitative and quantitative evaluation of the generated samples. Compatable with python 3.6 Data set net cifar 10

Generative-Adversarial-Network

How to Visualize Decision Trees using Matplotlib How to Visualize Decision Trees using Graphviz (what is Graphviz, how to install it on Mac and Windows, and how to use it to visualize decision trees) How to Visualize Individual Decision Trees from Bagged Trees or Random Forests

Visualizing-Decision-Trees-Random-forest-with-Python

This Project is for self learning Purpose

Multi-lable-classifiaction-on-tabluar-data-

Speech-to-text-using-deep-learning-

image classifiacion model with flask using keras

my-laptop-requirements-

Image-classification-model-deployment-using-Flask-

This is my first docker image this repo is for personal learning purpose

Docker-NLM-

Python-Challenges-

### Powerful EDA (Exploratory Data Analysis) using Sweetviz & pandas profiling

cat-and-dog

pandas-profiling-vs-Sweetviz-

HTML

Lungs-Segmentation

lungs segmentation without Mask image uisng Tradinational Computer vision (open-CV)

Converting-Image-into-base-64-and-coverting-base64-to-image

This Repo is for self learning purpose

Automated-image-classification

Yugen.ai

Yugen-ai-Heroku

Heroku-

this code is for self learning purpose. in this code we have deploy the a simple web app code with static file using flask on heroku server

This repo is for self learning Purpose

RASA-chat-bot-

Image-to-text-using-object-detection-

this repo is for self learning purpose

opencv-

This Repo is self learning purpose

Web-Scrapping-wiki

Machine Learning Flask CSS HTML Heroku Xgboost pycart

Nitinguptadu-Predictive-Tests-For-Assessing-Risk-of-Cancer-Recurrence-

Fetching data from website api and processing data and saving data in mongodb and saving data in csv and send mail using python on daily basis

scrapping-covid-19-Data-

Heroku-Demo

Covid-19-time-series-Classification

This Repo is for self learning purpose

Keras-cnn-for-Regressiion-and-classification--with-Multi-Target-

Ploting-of-Desion-Boundary-in-python-3.0

this repo is for self learning purpose

keras-ocr-

sending-email-using-python-

This repo is for self learning purpose for comparing two numerical columns and creating a new columns for saving higher have from two colomns

exp

Comaptaring-Two-columns-

Time-series-EEg-Classification

This repo is for self learning purpose . Saving models in a database and loading them using python is easy. We choose MongoDB because it is an open-source document database and leading NoSQL database.

HTML

Saving-ML-and-DL-models-in-MongoDB-using-python.

lasgger is a Flask extension to extract OpenAPI-Specification from all Flask views registered in your API. Flasgger also comes with SwaggerUI embedded so you can access http://localhost:5000/apidocs and visualize and interact with your API resources. Flasgger also provides validation of the incoming data, using the same specification it can validates if the data received as as a POST, PUT, PATCH is valid against the schema defined using YAML, Python dictionaries or Marshmallow Schemas. Flasgger can work with simple function views or MethodViews using docstring as specification, or using @swag_from decorator to get specification from YAML or dict and also provides SwaggerView which can use Marshmallow Schemas as specification. Flasgger is compatible with Flask-RESTful so you can use Resources and swag specifications together, take a look at restful example. Flasgger also supports Marshmallow APISpec as base template for specification, if you are using APISPec from Marshmallow take a look at apispec example.

Flasgger

Client: M17(https://m17.asia/en/product/17media/) a streaming platform. DataLink: Client_data.zip The attached data here is generated from the live-streaming platform. Analyze the data to come up with the top 20% streamers. Using these top 20% streamers as good streamers create a classification model which can classify whether any streamer is good streamer or not. Evaluation Metric : F1-Score Classification Report (Using sklearn.metrics.classification_report) is also required

Ankur-Machine-learning-

Use below directory structure, and generate text file containing (KITTI format ) annotation for each image. . ├── images │ ├── 000000.jpg │ . │ . │ └── XXXXXX.jpg └── annotations ├── 000000.txt . . └── XXXXXX.txt The text file should contain one annotation per line, with values separated by space. className 0.00 0 0.00 x1 y1 x2 y2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 className 0.00 0 0.00 x1 y1 x2 y2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Eg., person 0.00 0 0.00 20 50 196 600 0.00 0.00 0.00 0.00 0.00 0.00 0.00 face 0.00 0 0.00 100 120 150 180 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Note:- The dataset is available to download through Google drive & Kaggle , here's what can be done. Download it manually and upload to colab (https://www.kaggle.com/imneonizer/wider-person/download).

Object-detection-

Abstract: 100 recordings of a sensor array under different conditions in a home setting: background, wine and banana presentations. The array includes 8 MOX gas sensors, and humidity and temperature sensors. Source: Creators: Flavia Huerta,Gaurav Gawade Ramon Huerta, University of California San Diego, USA Donors: Flavia Huerta Ramon Huerta, University of California San Diego, USA (rhuerta â€˜@â€™ ucsd.edu) Thiago Mosqueiro, University of California San Diego, USA (thmosqueiro â€˜@â€™ ucsd.edu) Jordi Fonollosa, Institute for Bioengineering of Catalunya, Spain (jfonollosa â€˜@â€™ ibecbarcelona.eu) Nikolai Rulkov, University of California San Diego, USA ( nrulkov â€˜@â€™ ucsd.edu ) Irene Rodriguez-Lujan, Universidad Autonoma de Madrid, Spain ( Irene.rodriguez â€˜@â€™ uam.es ) Data Set Information: This dataset has recordings of a gas sensor array composed of 8 MOX gas sensors, and a temperature and humidity sensor. This sensor array was exposed to background home activity while subject to two different stimuli: wine and banana. The responses to banana and wine stimuli were recorded by placing the stimulus close to the sensors. The duration of each stimulation varied from 7min to 2h, with an average duration of 42min. This dataset contains a set of time series from three different conditions: wine, banana and background activity. There are 36 inductions with wine, 33 with banana and 31 recordings of background activity. One possible application is to discriminate among background, wine and banana. This dataset is composed of two files: HTsensordataset.dat (zipped), where the actual time series are stored, and the HTSensormetadata.dat, where metadata for each induction is stored. Each induction is uniquely identified by an id in both files. Thus, metadata for a particular induction can be easily found by matching columns id from each file. We also made available python scripts to exemplify how to import, organize and plot our data. The scripts are available on GitHub: https://github.com/gauravgawade951999/gauravgit For each induction, we include one hour of background activity prior to and after the stimulus presentation. Time series were recorded at one sample per second, with minor variations at some data points due to issues in the wireless communication. For details on which sensors were used and how the time series is organized, see Attribute Information below.

Gas-sensors-for-home-activity-monitoring-Data-Set